pith. sign in

arxiv: 2606.28338 · v1 · pith:APQR6EC6new · submitted 2026-05-30 · 💻 cs.IR

Memory Shot for Long-Term Dialogue

Chunyi Peng , Haidong Xin , Xuanshuo Sheng , Xin Dai , Zhenghao Liu , Shuo Wang , Yukun Yan , Zulong Chen
show 2 more authors
Yu Gu Ge Yu
This is my paper

Pith reviewed 2026-06-30 11:38 UTC · model grok-4.3

classification 💻 cs.IR
keywords long-term dialoguevisual memorymemory constructionepisode associationdialogue modelingLLM efficiency
0
0 comments X

The pith

MemShot renders local dialogue spans as visual units so models can link episodes across sessions using internal visual reasoning.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper sets out to replace text-heavy memory construction for long dialogues with a lighter visual approach. Existing systems extract and reorganize text evidence at high cost and lose cues such as speaker changes and turn order. MemShot instead turns short contiguous dialogue blocks into structured visual memory units that keep those cues intact. The model then uses its own visual reasoning to connect related episodes across time. This keeps performance steady on standard benchmarks while cutting memory-building time by a factor of seventy.

Core claim

MemShot renders local contiguous dialogue spans into structured visual memory units, preserving meta-information such as speaker transitions and turn boundaries, and relies on the model's internal visual reasoning capabilities to associate key episodes across sessions, avoiding the computational overhead of text-centered memory construction.

What carries the argument

Structured visual memory units created by rendering local contiguous dialogue spans, which carry chronological order and speaker meta-information for visual reasoning.

If this is right

  • MemShot matches prior methods on accuracy for LoCoMo and LongMemEval while shortening the memory pipeline.
  • It produces a 70 times speedup in memory construction.
  • Memory search shifts from surface lexical matching in flat text to structured local dialogue cues.
  • Speaker transitions and turn boundaries remain available inside each memory unit.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same visual-unit approach could be tested on sequential tasks outside dialogue, such as long document chains.
  • Performance may vary on models whose visual training is weaker than the ones evaluated here.
  • Adding color or layout variations to the visual renders might further strengthen episode separation.

Load-bearing premise

That the model's visual reasoning will reliably connect key episodes when presented with rendered visual dialogue units that retain speaker and turn structure.

What would settle it

A benchmark dataset of long dialogues with many cross-session references where the visual memory method retrieves fewer correct historical episodes than a text-based baseline.

Figures

Figures reproduced from arXiv: 2606.28338 by Chunyi Peng, Ge Yu, Haidong Xin, Shuo Wang, Xin Dai, Xuanshuo Sheng, Yu Gu, Yukun Yan, Zhenghao Liu, Zulong Chen.

Figure 1
Figure 1. Figure 1: Visualization of Memory Construction Latency and [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Illustration of Our Proposed MemShot. memory based on dialogue shots as memory units rather than fragmented and independent text chunks. 3.3 Efficient Memory Construction through Dialogue Chunk Shooting To more directly preserve the structural organization of raw dia￾logue, we introduce MemShot, a dialogue shooting mechanism that constructs structured visual memory units from local contiguous spans of the … view at source ↗
Figure 3
Figure 3. Figure 3: Performance of Text RAG and MemShot on the Lo [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 5
Figure 5. Figure 5: Memory-Augmented Generation of MLLMs Using [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Saliency Scores of the Input Evidence for Support [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Illustration of the Visual Memory Rendering Tem [PITH_FULL_IMAGE:figures/full_fig_p011_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Prompt used for LLM-as-a-Judge evaluation with GLM-5. [PITH_FULL_IMAGE:figures/full_fig_p013_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Case Study on Temporal Reasoning Scenario with Text RAG and MemShot. [PITH_FULL_IMAGE:figures/full_fig_p014_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Case Study on Multi-Session Evidence Aggregation Scenario with Text RAG and MemShot. [PITH_FULL_IMAGE:figures/full_fig_p015_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Prompt Template for Text RAG Inference Used in Our Experiments. [PITH_FULL_IMAGE:figures/full_fig_p016_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Prompt Template for MemShot Inference Used in Our Experiments. [PITH_FULL_IMAGE:figures/full_fig_p017_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Prompt Template for Rubric-Based Chain-of-Thought Analysis Used in Our Experiments. [PITH_FULL_IMAGE:figures/full_fig_p018_13.png] view at source ↗
read the original abstract

Large Language Models (LLMs) have demonstrated strong capabilities in general conversation, instruction following, and complex reasoning. However, in long-term dialogue settings, they often struggle to locate and utilize historical information most relevant to the current query. Existing approaches address this issue by constructing structured text-centered memory units through compressing and reorganizing user interaction history. However, these systems often rely on brute-force extraction of crucial evidence to associate episodes across dialogue sessions, causing substantial computational overhead and weakening structural cues such as speaker transitions, turn boundaries, and local contextual relationships. To avoid fragile text-based memory representations, we propose MemShot, which leverages dialogue structuring for long-term dialogue modeling and relies on the model's internal visual reasoning capabilities to associate key episodes. Specifically, MemShot renders local contiguous dialogue spans into structured visual memory units, preserving meta-information and chronological dialogue turns while avoiding heavy-weight textual memory construction. Experimental results show that MemShot achieves stable and competitive performance on both LoCoMo and LongMemEval, while substantially shortening the memory construction pipeline and delivering 70$\times$ speedup. Further analysis reveals that MemShot enhances the localization and utilization of historical evidence by directing memory processing toward structured local dialogue cues rather than surface-level lexical matching in a flat text stream. All codes are released on https://github.com/NEUIR/MemShot.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 2 minor

Summary. The paper proposes MemShot, a method for long-term dialogue that renders local contiguous dialogue spans into structured visual memory units. This leverages the LLM's internal visual reasoning to associate key episodes across sessions while preserving meta-information such as speaker transitions and turn boundaries. It claims to avoid the computational overhead of brute-force text extraction in prior structured memory approaches, achieving competitive results on the LoCoMo and LongMemEval benchmarks along with a 70× speedup in the memory construction pipeline. Code is released at the provided GitHub link.

Significance. If the empirical results hold, the work demonstrates a practical efficiency gain for memory-augmented long-term dialogue systems by shifting from text-centric to visual memory representations. The approach preserves structural dialogue cues that text compression often weakens and supplies reproducible code, which strengthens its utility for follow-on research in dialogue modeling.

major comments (1)
  1. The experimental claims of stable competitive performance rest on benchmark results whose presentation omits error bars, ablation studies isolating the visual reasoning component, and basic dataset statistics for LoCoMo and LongMemEval; these omissions make it difficult to assess whether the reported gains are robust or attributable to the proposed visual memory mechanism rather than other pipeline choices.
minor comments (2)
  1. The abstract states that MemShot 'directs memory processing toward structured local dialogue cues rather than surface-level lexical matching'; a brief concrete example of this distinction in the main text would clarify the claimed advantage over prior text-based methods.
  2. The phrase 'All codes are released' should be revised to 'The code is released' for grammatical consistency.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive feedback and recommendation for minor revision. We address the single major comment below.

read point-by-point responses

  1. Referee: The experimental claims of stable competitive performance rest on benchmark results whose presentation omits error bars, ablation studies isolating the visual reasoning component, and basic dataset statistics for LoCoMo and LongMemEval; these omissions make it difficult to assess whether the reported gains are robust or attributable to the proposed visual memory mechanism rather than other pipeline choices.

    Authors: We agree that the original presentation omitted these elements. In the revision we will add basic dataset statistics for LoCoMo and LongMemEval. We will also report error bars (standard deviation across repeated runs) to support the claim of stable performance. Our existing comparisons against text-centered structured memory baselines already isolate the contribution of the visual representation; a dedicated ablation study focused solely on the visual reasoning component would require new experiments that exceed the scope of a minor revision, but we can expand the discussion of the existing comparisons if space allows. These additions will strengthen the manuscript without changing the core claims or the reported 70× speedup. revision: partial

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper introduces MemShot as an empirical method that renders local dialogue spans into structured visual memory units to leverage LLM visual reasoning for episode association, with claims resting on benchmark results (LoCoMo, LongMemEval) and measured 70× speedup from pipeline shortening. No equations, fitted parameters, predictions, or derivation chain exist that could reduce to self-defined inputs or self-citations. The approach is presented directly via implementation details and experimental protocols without invoking load-bearing self-citations, uniqueness theorems, or ansatzes; the central performance claims are externally falsifiable via the released code and benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the domain assumption that LLMs possess reliable internal visual reasoning for associating dialogue episodes from rendered images; no free parameters or invented entities are introduced in the abstract.

axioms (1)
  • domain assumption LLMs have internal visual reasoning capabilities sufficient to associate key episodes from rendered dialogue images
    Invoked as the core mechanism enabling the approach (abstract: 'relies on the model's internal visual reasoning capabilities to associate key episodes')

pith-pipeline@v0.9.1-grok · 5788 in / 1186 out tokens · 24999 ms · 2026-06-30T11:38:20.347221+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

49 extracted references · 30 canonical work pages · 17 internal anchors

  1. [1]

    Shuai Bai, Yuxuan Cai, Ruizhe Chen, Keqin Chen, Xionghui Chen, Zesen Cheng, Lianghao Deng, Wei Ding, Chang Gao, Chunjiang Ge, et al . 2025. Qwen3-vl technical report.ArXiv preprintabs/2511.21631 (2025). https://arxiv.org/abs/ 2511.21631

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  2. [2]

    Sinchana Ramakanth Bhat, Max Rudat, Jannis Spiekermann, and Nicolas Flores- Herr. 2025. Rethinking chunk size for long-document retrieval: A multi-dataset analysis.ArXiv preprintabs/2505.21700 (2025). https://arxiv.org/abs/2505.21700

    work page arXiv 2025

  3. [3]

    Jiale Cheng, Yusen Liu, Xinyu Zhang, Yulin Fei, Wenyi Hong, Ruiliang Lyu, Weihan Wang, Zhe Su, Xiaotao Gu, Xiao Liu, et al. 2025. Glyph: Scaling context windows via visual-text compression.ArXiv preprintabs/2510.17800 (2025). https://arxiv.org/abs/2510.17800

    work page arXiv 2025

  4. [4]

    Prateek Chhikara, Dev Khant, Saket Aryan, Taranjeet Singh, and Deshraj Ya- dav. 2025. Mem0: Building production-ready ai agents with scalable long-term memory.ArXiv preprintabs/2504.19413 (2025). https://arxiv.org/abs/2504.19413

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  5. [5]

    Yiming Du, Wenyu Huang, Danna Zheng, Zhaowei Wang, Sebastien Montella, Mirella Lapata, Kam-Fai Wong, and Jeff Z Pan. 2025. Rethinking Memory in LLM based Agents: Representations, Operations, and Emerging Topics.ArXiv preprint abs/2505.00675 (2025). https://arxiv.org/abs/2505.00675

    work page arXiv 2025

  6. [6]

    Sinan Fan, Liang Xie, Chen Shen, Ge Teng, Xiaosong Yuan, Xiaofeng Zhang, Chenxi Huang, Wenxiao Wang, Xiaofei He, and Jieping Ye. 2025. Improving complex reasoning with dynamic prompt corruption: A soft prompt optimization approach.ArXiv preprintabs/2503.13208 (2025). https://arxiv.org/abs/2503.13208

    work page arXiv 2025

  7. [7]

    Jizhan Fang, Xinle Deng, Haoming Xu, Ziyan Jiang, Yuqi Tang, Ziwen Xu, Shumin Deng, Yunzhi Yao, Mengru Wang, Shuofei Qiao, et al . 2025. Light- mem: Lightweight and efficient memory-augmented generation.ArXiv preprint abs/2510.18866 (2025). https://arxiv.org/abs/2510.18866

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  8. [8]

    Michael Günther, Saba Sturua, Mohammad Kalim Akram, Isabelle Mohr, An- drei Ungureanu, Bo Wang, Sedigheh Eslami, Scott Martens, Maximilian Werk, Nan Wang, et al. 2025. jina-embeddings-v4: Universal embeddings for multi- modal multilingual retrieval. InProceedings of the 5th Workshop on Multilingual Representation Learning (MRL 2025). 531–550

    2025

  9. [9]

    Helia Hashemi, Jason Eisner, Corby Rosset, Benjamin Van Durme, and Chris Kedzie. 2024. Llm-rubric: A multidimensional, calibrated approach to automated evaluation of natural language texts. InProceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 13806– 13834

    2024

  10. [10]

    Demis Hassabis and Eleanor A Maguire. 2007. Deconstructing episodic memory with construction.Trends in cognitive sciences11, 7 (2007), 299–306

    2007

  11. [11]

    Chuanrui Hu, Xingze Gao, Zuyi Zhou, Dannong Xu, Yi Bai, Xintong Li, Hui Zhang, Tong Li, Chong Zhang, Lidong Bing, et al . 2026. EverMemOS: A Self- Organizing Memory Operating System for Structured Long-Horizon Reasoning. ArXiv preprintabs/2601.02163 (2026). https://arxiv.org/abs/2601.02163

    work page arXiv 2026

  12. [12]

    Yuyang Hu, Shichun Liu, Yanwei Yue, Guibin Zhang, Boyang Liu, Fangyi Zhu, Jiahang Lin, Honglin Guo, Shihan Dou, Zhiheng Xi, et al. 2025. Memory in the age of ai agents.ArXiv preprintabs/2512.13564 (2025). https://arxiv.org/abs/ 2512.13564

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  13. [13]

    Dongming Jiang, Yi Li, Songtao Wei, Jinxin Yang, Ayushi Kishore, Alysa Zhao, Dingyi Kang, Xu Hu, Feng Chen, Qiannan Li, et al. 2026. Anatomy of Agentic Memory: Taxonomy and Empirical Analysis of Evaluation and System Limita- tions.ArXiv preprintabs/2602.19320 (2026). https://arxiv.org/abs/2602.19320

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  14. [14]

    Jiazheng Kang, Mingming Ji, Zhe Zhao, and Ting Bai. 2025. Memory os of ai agent. InProceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 25972–25981

    2025

  15. [15]

    Patrick AF Laing and Joseph E Dunsmoor. 2025. Event segmentation promotes the reorganization of emotional memory.Journal of Cognitive Neuroscience37, 1 (2025), 110–134

    2025

  16. [16]

    Patrick S. H. Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, and Douwe Kiela. 2020. Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks. InAdvances in Neural In- formation Processing Systems 33: Annual Conference on Neural Inf...

    2020

  17. [17]

    Mingxin Li, Yanzhao Zhang, Dingkun Long, Keqin Chen, Sibo Song, Shuai Bai, Zhibo Yang, Pengjun Xie, An Yang, Dayiheng Liu, et al . 2026. Qwen3-VL- Embedding and Qwen3-VL-Reranker: A Unified Framework for State-of-the- Art Multimodal Retrieval and Ranking.ArXiv preprintabs/2601.04720 (2026). https://arxiv.org/abs/2601.04720

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  18. [18]

    Yanhong Li, Zixuan Lan, and Jiawei Zhou. 2025. Text or Pixels? Evaluating Efficiency and Understanding of LLMs with Visual Text Inputs. InFindings of the Association for Computational Linguistics: EMNLP 2025. 10564–10578

    2025

  19. [19]

    Zhiyu Li, Chenyang Xi, Chunyu Li, Ding Chen, Boyu Chen, Shichao Song, Simin Niu, Hanyu Wang, Jiawei Yang, Chen Tang, et al. 2025. Memos: A memory os for ai system.ArXiv preprintabs/2507.03724 (2025). https://arxiv.org/abs/2507.03724

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  20. [20]

    Jiaheng Liu, Dawei Zhu, Zhiqi Bai, Yancheng He, Huanxuan Liao, Haoran Que, Zekun Wang, Chenchen Zhang, Ge Zhang, Jiebin Zhang, et al. 2025. A compre- hensive survey on long context language modeling.ArXiv preprintabs/2503.17407 (2025). https://arxiv.org/abs/2503.17407

    work page arXiv 2025

  21. [21]

    Liu, Kevin Lin, John Hewitt, Ashwin Paranjape, Michele Bevilacqua, Fabio Petroni, and Percy Liang

    Nelson F. Liu, Kevin Lin, John Hewitt, Ashwin Paranjape, Michele Bevilacqua, Fabio Petroni, and Percy Liang. 2024. Lost in the Middle: How Language Models Use Long Contexts.Transactions of the Association for Computational Linguistics 12 (2024), 157–173. doi:10.1162/tacl_a_00638

    work page doi:10.1162/tacl_a_00638 2024

  22. [22]

    Zhenghao Liu, Pengcheng Huang, Zhipeng Xu, Xinze Li, Shuliang Liu, Chunyi Peng, Haidong Xin, Yukun Yan, Shuo Wang, Xu Han, et al . 2026. Knowledge intensive agents.AI Open(2026)

    2026

  23. [23]

    Yujie Lu, Xiujun Li, Tsu-Jui Fu, Miguel Eckstein, and William Yang Wang. 2024. From text to pixel: Advancing long-context understanding in mllms.ArXiv preprintabs/2405.14213 (2024). https://arxiv.org/abs/2405.14213

    work page arXiv 2024

  24. [24]

    Adyasha Maharana, Dong-Ho Lee, Sergey Tulyakov, Mohit Bansal, Francesco Barbieri, and Yuwei Fang. 2024. Evaluating very long-term conversational mem- ory of llm agents. InProceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 13851–13870

    2024

  25. [25]

    Yufan Mao, Hanjing Ye, Wenlong Dong, Chengjie Zhang, and Hong Zhang. 2025. Meta-Memory: Retrieving and Integrating Semantic-Spatial Memories for Robot Conference’17, July 2017, Washington, DC, USA Chunyi Peng et al. Spatial Reasoning.ArXiv preprintabs/2509.20754 (2025). https://arxiv.org/abs/ 2509.20754

    work page arXiv 2025

  26. [26]

    Sophie Nolden, Gözem Turan, Berna Güler, and Eren Günseli. 2024. Prediction error and event segmentation in episodic memory.Neuroscience & Biobehavioral Reviews157 (2024), 105533

    2024

  27. [27]

    Charles Packer, Vivian Fang, Shishir_G Patil, Kevin Lin, Sarah Wooders, and Joseph_E Gonzalez. 2023. MemGPT: towards LLMs as operating systems. (2023)

    2023

  28. [28]

    Ori Ram, Yoav Levine, Itay Dalmedigos, Dor Muhlgay, Amnon Shashua, Kevin Leyton-Brown, and Yoav Shoham. 2023. In-Context Retrieval-Augmented Lan- guage Models.Transactions of the Association for Computational Linguistics11 (2023), 1316–1331. doi:10.1162/tacl_a_00605

    work page doi:10.1162/tacl_a_00605 2023

  29. [29]

    Preston Rasmussen, Pavlo Paliychuk, Travis Beauvais, Jack Ryan, and Daniel Chalef. 2025. Zep: a temporal knowledge graph architecture for agent memory. ArXiv preprintabs/2501.13956 (2025). https://arxiv.org/abs/2501.13956

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  30. [30]

    2026.Seed 2.0 Model Card: Towards Intelligence Frontier for Real- World Complexity

    ByteDance Seed. 2026.Seed 2.0 Model Card: Towards Intelligence Frontier for Real- World Complexity. Technical Report. Technical report (model card), February

    2026

  31. [31]

    URL https://lf3-static

  32. [32]

    Yaorui Shi, Shugui Liu, Yu Yang, Wenyu Mao, Yuxin Chen, Qi Gu, Hui Su, Xunliang Cai, Xiang Wang, and An Zhang. 2026. MemOCR: Layout-Aware Visual Memory for Efficient Long-Horizon Reasoning.ArXiv preprintabs/2601.21468 (2026). https://arxiv.org/abs/2601.21468

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  33. [33]

    Zhen Tan, Jun Yan, I-Hung Hsu, Rujun Han, Zifeng Wang, Long Le, Yiwen Song, Yanfei Chen, Hamid Palangi, George Lee, et al. 2025. In prospect and retrospect: Reflective memory management for long-term personalized dialogue agents. InProceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 8416–8439

    2025

  34. [34]

    Alex Jinpeng Wang, Linjie Li, Yiqi Lin, Min Li, Lijuan Wang, and Mike Zheng Shou. 2024. Leveraging Visual Tokens for Extended Text Contexts in Multi- Modal Learning. InAdvances in Neural Information Processing Systems 38: Annual Conference on Neural Information Processing Systems 2024, NeurIPS 2024, Vancouver, BC, Canada, December 10 - 15, 2024, Amir Glob...

    2024

  35. [35]

    Weiyun Wang, Zhangwei Gao, Lixin Gu, Hengjun Pu, Long Cui, Xingguang Wei, Zhaoyang Liu, Linglin Jing, Shenglong Ye, Jie Shao, et al. 2025. Internvl3. 5: Ad- vancing open-source multimodal models in versatility, reasoning, and efficiency. ArXiv preprintabs/2508.18265 (2025). https://arxiv.org/abs/2508.18265

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  36. [36]

    Xindi Wang, Mahsa Salmani, Parsa Omidi, Xiangyu Ren, Mehdi Rezagholizadeh, and Armaghan Eshaghi. 2024. Beyond the Limits: A Survey of Techniques to Extend the Context Length in Large Language Models. InProceedings of the Thirty-Third International Joint Conference on Artificial Intelligence, IJCAI 2024, Jeju, South Korea, August 3-9, 2024. ijcai.org, 8299...

    2024

  37. [37]

    Haoran Wei, Yaofeng Sun, and Yukun Li. 2025. Deepseek-ocr: Contexts optical compression.ArXiv preprintabs/2510.18234 (2025). https://arxiv.org/abs/2510. 18234

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  38. [38]

    Chi, Quoc V

    Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Brian Ichter, Fei Xia, Ed H. Chi, Quoc V. Le, and Denny Zhou. 2022. Chain-of-Thought Prompt- ing Elicits Reasoning in Large Language Models. InAdvances in Neural Infor- mation Processing Systems 35: Annual Conference on Neural Information Pro- cessing Systems 2022, NeurIPS 2022, New Orleans, LA, USA,...

    2022

  39. [39]

    Di Wu, Hongwei Wang, Wenhao Yu, Yuwei Zhang, Kai-Wei Chang, and Dong Yu. 2024. Longmemeval: Benchmarking chat assistants on long-term interactive memory.ArXiv preprintabs/2410.10813 (2024). https://arxiv.org/abs/2410.10813

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  40. [40]

    Yaxiong Wu, Yongyue Zhang, Sheng Liang, and Yong Liu. 2025. Sgmem: Sen- tence graph memory for long-term conversational agents.ArXiv preprint abs/2509.21212 (2025). https://arxiv.org/abs/2509.21212

    work page arXiv 2025

  41. [41]

    Haidong Xin, Xinze Li, Zhenghao Liu, Yukun Yan, Shuo Wang, Cheng Yang, Yu Gu, Ge Yu, and Maosong Sun. 2026. MetaMem: Evolving Meta-Memory for Knowledge Utilization through Self-Reflective Symbolic Optimization.ArXiv preprintabs/2602.11182 (2026). https://arxiv.org/abs/2602.11182

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  42. [42]

    Wujiang Xu, Zujie Liang, Kai Mei, Hang Gao, Juntao Tan, and Yongfeng Zhang

  43. [43]

    A-MEM: Agentic Memory for LLM Agents

    A-mem: Agentic memory for llm agents.ArXiv preprintabs/2502.12110 (2025). https://arxiv.org/abs/2502.12110

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  44. [44]

    An Yang, Anfeng Li, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Gao, Chengen Huang, Chenxu Lv, et al. 2025. Qwen3 technical report.ArXiv preprintabs/2505.09388 (2025). https://arxiv.org/abs/2505.09388

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  45. [45]

    Tianyu Yu, Zefan Wang, Chongyi Wang, Fuwei Huang, Wenshuo Ma, Zhihui He, Tianchi Cai, Weize Chen, Yuxiang Huang, Yuanqian Zhao, et al. 2025. Minicpm-v 4.5: Cooking efficient mllms via architecture, data, and training recipe.ArXiv preprintabs/2509.18154 (2025). https://arxiv.org/abs/2509.18154

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  46. [46]

    Peter Zeidman, Sinéad L Mullally, and Eleanor A Maguire. 2015. Construct- ing, perceiving, and maintaining scenes: hippocampal activity and connectivity. Cerebral Cortex25, 10 (2015), 3836–3855

    2015

  47. [47]

    Aohan Zeng, Xin Lv, Zhenyu Hou, Zhengxiao Du, Qinkai Zheng, Bin Chen, Da Yin, Chendi Ge, Chengxing Xie, Cunxiang Wang, et al. 2026. GLM-5: from Vibe Coding to Agentic Engineering.ArXiv preprintabs/2602.15763 (2026). https: //arxiv.org/abs/2602.15763

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  48. [48]

    Wanjun Zhong, Lianghong Guo, Qiqi Gao, He Ye, and Yanlin Wang. 2024. Memo- ryBank: Enhancing Large Language Models with Long-Term Memory. InThirty- Eighth AAAI Conference on Artificial Intelligence, AAAI 2024, Thirty-Sixth Con- ference on Innovative Applications of Artificial Intelligence, IAAI 2024, Fourteenth Symposium on Educational Advances in Artific...

    work page doi:10.1609/aaai.v38i17.29946 2024

  49. [49]

    next month

    Zihan Zhou, Chong Li, Xinyi Chen, Shuo Wang, Yu Chao, Zhili Li, Haoyu Wang, Qi Shi, Zhixing Tan, Xu Han, et al. 2025. LLM × MapReduce: Simplified Long- Sequence Processing using Large Language Models. InProceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 27664–27678. Memory Shot for Long-Term Di...

    work page arXiv 2025